US 5668481 A Abstract A sequence generator for generating multiple-pattern sequences utilizing an inverting non-linear autonomous machine is disclosed. The generation of the sequence is based on the interdependency relationship between the bits of the deterministic patterns in the sequences to be generated. The autonomous machine comprises a number of flip-flop cascades each containing a number of flip-flops. The flip-flops in each of the cascades are connected in a way that the output of one flip-flop is connected to the input of the next flip-flop in each cascade. The autonomous machine further comprises a number of XOR gates each feeding its output to the input of a corresponding one of the cascades of flip-flops. The autonomous machine further comprises a switch device that includes a number of switches for receiving the output of each of the flip-flops in the cascades as feedbacks for outputting to the inputs of the XOR gates. Each of the XOR gates is selectively added with an inverter for inverting its output before connecting to the corresponding one of the cascade of flip-flops. Each of the switches in the switch device is selectively programmed for connecting corresponding one of the feedbacks to the input of the corresponding one of the XOR gates.
Claims(6) 1. A process for generating multiple-pattern sequences utilizing an inverting non-linear autonomous machine based on the interdependency relationship between M deterministic patterns each having a width of N bits in said sequences, said autonomous machine comprising a plurality of flip-flop cascades each containing a plurality of said flip-flops, a plurality of XOR gates each feeding output thereof to the input of a corresponding cascade of flip-flops, and a switch means including a plurality of switches for receiving each output of said flip-flops in said cascades as feedbacks for outputting to the inputs of said plurality of XOR gates, said process comprising the steps of:
(a) setting up a set of dependent equations describing the interdependency between all bits in the width of each of said deterministic patterns utilizing a set of variables including the inversion of the bits in said patterns, and the interdependency of the bits in said patterns; (b) resolving said set of dependent equations into a unique set of solutions expressed by said inversion and interdependency variables; and (c) said inversion and interdependency variable solutions being employed for the selection of addition of inverters to outputs of said XOR gates, and the selection of programming of said switches in said switch means respectively. 2. The process for generating multiple-pattern sequences of claim 1, wherein step (a) comprises implementing a time-frame expansion scheme for setting up said set of dependent equations.
3. The process for generating multiple-pattern sequences of claim 1, wherein step (b) comprises implementing a binary relaxation scheme for resolving said solution of said set of dependent equations.
4. A sequence generator for generating multiple-pattern sequences utilizing an inverting non-linear autonomous machine based on the interdependency relationship between M deterministic patterns each having a width of N bits in said sequences, said autonomous machine comprising:
a plurality of flip-flop cascades each containing a plurality of flip-flops, said flip-flops in each of said cascades being connected with the output of one flip-flop connected to the input of the next flip-flop in said each cascade; a plurality of XOR gates each feeding output thereof to the input of a corresponding one of said cascades of flip-flops; and a switch means including a plurality of switches for receiving the output of each of said flip-flops in said cascades as feedbacks for outputting to the inputs of said plurality of XOR gates; wherein each of said XOR gates being selectively added with an inverter for inverting output thereof before connecting to said corresponding one of said cascade of flip-flops, and each of said switches in said switch means being selectively programmed for connecting corresponding one of said feedbacks to the input of the corresponding one of said XOR gates. 5. The sequence generator for generating multiple-pattern sequences of claim 4, wherein each of said cascades of flip-flops contains at least one flip-flop.
6. The sequence generator for generating multiple-pattern sequences of claim 4, wherein said switch means comprises a grid-work of programmable switches.
Description 1. Field of the Invention The present invention relates in general to multiple-pattern sequence generators based on inverting non-linear autonomous machines. In particular, the present invention relates to multiple-pattern sequence generators based on inverting non-linear autonomous machines, capable of relying on the interdependency between deterministic patterns in a sequence for generating a multiple of patterns in a sequence containing the leading deterministic patterns and followed by pseudo-random patterns. More particularly, the present invention relates to multiple-pattern sequence generators based on inverting non-linear autonomous machines having simplified hardware structural configuration. 2. Technical Background Sequence generators are widely utilized in digital electronic equipments such as controllers, coders, decoders, and test-pattern generators etc. The principal hardware structural configuration of some of the conventional sequence generators are based on digital counters which enjoy the primary advantage of simplicity in their electronic circuitries. However, they suffer the major limitation in the generation of fixed sequences. Others of the conventional sequence generator employ finite state machines for the description of the interdependent relationship between the sequences and utilize random logic to construct the circuitry for the generation of the desired sequences. The advantage of this implementation is the ability to generate any specific sequence, but with the obvious disadvantage of the lack of regularity in the constructed circuitries. Such circuitries are also inadequate for the generation of pseudo-random sequences. Still other conventional sequence generators employ the category of linear feedback shift registers (LFSR) as the primary core for the generation of pseudo-random sequences. Maximum lengths of pseudo-random sequences can be advantageously generated with simple and regular hardware structural configurations. However, the generated sequences are not in arbitrary order, but are restricted by the feedback path in the LFSR's. Specific deterministic patterns in the generated sequences for practical applications are frequently required to appear as early as possible, or, they may also be required to appear repeatedly, or in specific order in the generated pseudo-random sequences. These conditions are not totally satisfiable with the conventional state-of-the-art and straight-forward LFSR's. With the LFSR being widely employed in the generation of pseudo-random sequences having maximized lengths, reseeding practices of the LFSR's are utilized to change the order of the patterns in the generated sequences, causing the desired pattern to appear as early as possible. Alteration of the feedback paths in the LFSR's can result in the change of the appearing order of the patterns in the generated sequences, thereby generating some of the required specific patterns. However, unless certain stored-pattern procedures are taken to pre-store ordered sequences or specific sequences in memory devices before LFSR's can be used to generate the pseudo-random sequence, current LFSR techniques can only satisfy some of the practical application requirements. Otherwise, instead of simplicity and regularity, huge and complex hardware configuration will be necessary for the generation of the practically arbitrary sequences of repeated and pseudo-random nature. One important application of multiple-pattern sequence generator is in the testing of very large-scale integrated circuit (VLSI) devices. VLSI technology is advancing rapidly, electronic circuitries with complex functionality and huge scale of integration are being integrated into single IC packages. These VLSI IC packages do not have sufficient lead pins for the arrangement of the connection to external testing equipment for the IC device testing purposes. Internal testing circuitries are therefore included as a portion of the entire IC device for conducting on-board testing from inside the VLSI IC device. This built-in self-test (BIST) technique has being widely accepted not only because it reduces the cost and difficulty of the external testing practice, but also because of its ability to provide the real-time on-line testing of the VLSI device. In a typical BIST design, a sequence generator is required for the generation of a test sequence containing necessary testing patterns. As is in the case of the above-discussed sequence generators, most BIST sequence generator designs employ LFSR's as the basis for the generation of the test sequences. Several different techniques have been developed for the construction of the test sequence generators of the BIST designs. For example, in the exhaustive testing scheme, all possible input combinations are generated as the test pattern of the BIST-based sequence generator. This scheme, however, is not suitable for circuitries with requiring large amount of inputs. In another pseudo-exhaustive testing scheme, the electronic circuitry of the BIST test device is partitioned into a number of smaller sub-circuits. Each of the subcircuits performs its own exhaustive testing procedure as in the exhaustive testing scheme described above. This pseudo-exhaustive testing scheme is limited by the fact that it can not generate deterministic sequences or sequences with repeated test patterns. Such limitation prohibits the use of this pseudo-exhaustive testing scheme in sequential circuits or testing schemes concerning delay faults. In another pseudo-random testing scheme, a pseudo-random pattern generator is utilized to generate the test patterns in sequences. This scheme, however, is restricted in the detection of some random pattern resistant faults, and is neither suitable for use in sequential circuits nor in the testing scheme concerning delay faults. In still another mixed-mode testing scheme, pseudo-random test patterns are generated first and then certain test patterns of specific arrangement. The hardware structural configuration of this testing scheme, however, requires much more complication. It is therefore an object of the present invention to provide a process for generating multiple-pattern sequences employing inverting non-linear autonomous machines based on the interdependency between deterministic patterns in the sequence. It is another object of the present invention to provide a process for generating multiple-pattern sequences containing deterministic patterns followed by pseudo-random patterns employing inverting non-linear autonomous machines based on the interdependency between deterministic patterns in the sequence. It is still therefore an object of the present invention to provide a multiple-pattern sequence generator based on an inverting non-linear autonomous machine capable of relying on the interdependency between deterministic patterns in a sequence for generating a multiple of patterns in a sequence containing the leading deterministic patterns and followed by pseudo-random patterns. It is yet another object of the present invention to provide a multiple-pattern sequence generator based on an inverting non-linear autonomous machine that has the simplest possible hardware structural configuration. The present invention achieves the above identified object by providing a process for generating multiple-pattern sequences utilizing an inverting non-linear autonomous machine based on the interdependency relationship between the bits of the deterministic patterns in the sequences to be generated. The process comprises the steps of (a), setting up a set of dependent equations describing the interdependency between all bits in the width of each of the deterministic patterns utilizing a set of variables that includes the inversion of the bits in the patterns, and the interdependency of the bits in the patterns; (b), resolving the set of dependent equations into a unique set of solutions expressed by the inversion and interdependency variables; and (c), the inversion and interdependency variable solutions are employed for the selection of addition of inverters to outputs of exclusive-OR (XOR) gates, and the selection of programming of the switches in the switch means respectively. The present invention achieves the above identified object also by providing a sequence generator for generating multiple-pattern sequences utilizing an inverting non-linear autonomous machine. The generation of the sequence is based on the interdependency relationship between the bits of the deterministic patterns in the sequences to be generated. The autonomous machine comprises a number of flip-flop cascades each containing a number of flip-flops. The flip-flops in each of the cascades are connected in a way that the output of one flip-flop is connected to the input of the next flip-flop in each cascade. The autonomous machine further comprises a number of XOR gates each feeding its output to the input of a corresponding one of the cascades of flip-flops. The autonomous machine further comprises a switch device that includes a number of switches for receiving the output of each of the flip-flops in the cascades as feedbacks for outputting to the inputs of the XOR gates. Each of the XOR gates is selectively added with an inverter for inverting its output before connecting to the corresponding one of the cascade of flip-flops. Each of the switches in the switch device is selectively programmed for connecting corresponding one of the feedbacks to the input of the corresponding one of the XOR gates. Other objects, features and advantages of the present invention will become apparent by way of the following detailed description of the preferred but non-limiting embodiment. The description is made with reference to the accompanied drawings, wherein: FIG. 1 shows the schematic diagram of a general multiple-pattern sequence generator based on an inverting non-linear autonomous machine in accordance with a preferred embodiment of the present invention; and FIG. 2 shows the schematic diagram of a specific multiple-pattern sequence generator based on an inverting non-linear autonomous machine in accordance with an exemplified outline of a deterministic pattern designation. Refer to FIG. 1 of the drawing. FIG. 1 shows the schematic diagram of a multiple-pattern sequence generator based on the inverting non-linear autonomous machine in accordance with a preferred embodiment of the present invention. The multiple-pattern sequence generator of the present invention, as is shown, comprises a number of feedback shift registers, FSR's 33, a number of inverted and non-inverted XOR gates 44, and a switch device 55. Each of the FSR's 33 may comprise an inverting non-linear two-dimensional feedback shift registers. For example, the FSR 33 can be consisted of at least one flip-flop 333. In the case of more than one flip-flops 333, all the flip-flops 333 are connected in cascade, each having its output connected to both the input of the flip-flop in the next stage and back to the switch device 55. The input of each of the FSR's 33, namely the input of the first stage of the flip-flop 333 in the cascade, is fed by the output of a corresponding one of the number of inverted/non-inverted XOR gates 44. Each of the inverted/non-inverted XOR gates 44 may be a linear XOR gate 44a, or a linear XOR gate combined by a non-linear inverter 44b, as is shown in FIG. 2, which is to be described later. There is a corresponding number of these inverted/non-inverted XOR gates 44 as there is the number of FSR's 33, and, as is mentioned above, the output of each of these inverted/non-inverted XOR gates 44 is fed to the input of the correspondingly connected FSR 33. The input to each of these inverted/non-inverted XOR gates 44 is, as can be observed in the drawing, connected to the switch device 55. The switch device 55, which feeds each inputs of the inverted/non-inverted XOR gates 44 and also receives feedbacks from the constituent stages of flip-flops in the FSR's 33, is shown to comprise a grid-work of programmable switches 555. The purpose of the utilization of the switch device 55 is to selectively send the feedback signals of the stages of flip-flop flops in the FSR's 33 back into the inputs of the inverted/non-inverted XOR gates 44 again. The selection of the feedback paths in the switch device 55, which is to be described later in more detail, is based on the interdependency between the deterministic patterns in the sequence to be generated. The multiple-pattern sequence generator as shown in FIG. 1 generally outlines the hardware structural configuration of a device in accordance with the disclosure of the present invention that can be utilized for the generation of the desired sequences. The object of the present invention, as outlined above, to provide a multiple-pattern sequence generator, capable of implementing the process of the present invention to rely on the interdependency between deterministic patterns in a sequence for generating a multiple of patterns in the sequence containing the leading deterministic patterns and followed by pseudo-random patterns, can be described in other terms. That is, for a given sequence of patterns having a width of N bits and a length of M patterns, a multiple-pattern sequence generator is to be designed based on an inverting non-linear autonomous machine, so as to be able to generate a sequence, containing the first M patterns completely in conformity to the deterministic patterns in the sequence, and followed by the pseudo-random patterns. Moreover, the hardware is required to be of the simplest possible configuration, making use of least possible XOR gates, inverters, and flip-flops. Thus, in the above-outlined problem, the total of N single-bits in each of the patterns of the sequence can be considered to be independent signals, and may each be generated utilizing one of a total of N independent signal generators, for example, N LFSR's. Basically, it is possible to seek the linear dependency, or inverse linear dependency, between the signals in the patterns. Based on the resolved dependency relationships, it is possible to simplify the configuration of the hardware structure used for the generation of the sequence of patterns. A hardware structure for the multiple-pattern sequence generator requiring the minimum number of XOR gates, inverters, and flip-flops can then be achieved. A simple example is described below to show that the above-mentioned scheme of seeking the interdependency relationship can substantially be employed to build a sequence generator that is capable of generating a pseudo-random sequence containing the necessary deterministic patterns. For the specified deterministic patterns, a series of shift registers having sufficient length (that is, the length of the deterministic sequence) can be employed to store the entire sequence therein. The shift registers can constitute a random LFSR that is able to generate the pseudo-random patterns followed thereafter by the deterministic patterns. With this minimum length of shift registers, the worst case condition of the multiple-pattern sequence generator can still be met. In the process of seeking the interdependency of the deterministic patterns in the sequence, a technique of time-frame expansion is employed to gradually increase the length of the cascade of the shift registers in the sequence generator hardware, until the ultimate shift register cascade having the shortest possible length is resolved. The cascade length is increased based on the required delay signals. That is, the shift registers serve to store the individual signals in delay in the cascaded length of the registers, and, further, the delayed signals in the registers can in turn be used again as the new source of incoming signals. When the inverted versions of the signals, as well as of the delayed signals are added into the manipulation of the scheme, the space in which the interdependency is sought is greatly expanded. Due to the fact that inverting non-linear shift registers allow only the incorporation of linear XOR gate (as well as its inverted version), the above-outlined process of seeking the ultimate shift register cascade can therefore be converted into the problem of resolving non-invertedly signaled binary linear dependent equations and invertedly signaled binary linear dependent equations. The length of the shift register cascade is a parameter dependent on the scale of the dependent equations. The longer the length of the shift register cascade, the more the number of independent signals, and the easier the resolution of the linear dependent equations will become. Therefore, with the resolved interdependency between each of the signals in a pattern and with the delayed signals, the length of the shift register cascade for the sequence generator can be greatly reduced. This is expressed in the following equations: For a sequence having a length of M patterns and a width of N, if V Then, for each V
V and
a The condition of a In the equation outlined above, the number m signifies the time-frame expansion index. In the process of a total of m expansions, for every V To show the practical feasibility of the above-described scheme of finding the hardware combination for the sequence generator, portions of a set of computer program algorithm implementing this scheme has been included herein for reference. The exemplified algorithm serves to read in the desired patterns in the sequence to be generated and produces the dependent equations of equation (1) listed above, with inverted and non-inverted signal variables. The time-frame expansion scheme as described above is then implemented in the exemplified algorithm to seek to resolve into the preferred set of solutions. Due to the fact that the solution to the dependent equations (1) is not unique, the solution produced by the algorithm, based on the establishment of some guide rules, may be set to be optimized at, for example, a solution with a minimum number of flip-flops required, the solution with a minimum feedback, a solution with a minimum number of XOR gates required, and/or the greatest extent in the generated pseudo-random distribution. Algorithm Listing 1 Specifically, Algorithm Listing 1 shows the algorithm employing the time-frame expansion scheme for the generation of the dependent equations (1) that describe a targeted deterministic patterns characterized by the essential initial patterns:
______________________________________Find Listing 2 Algorithm Listing 2, on the other hand, shows the algorithm employing a binary relaxation scheme for the solution of the dependent equations:
______________________________________Solve A specific example of the procedure of pursuing the mathematical solution to the dependent equations for the construction of the multiple-pattern sequence generator in accordance with the present invention is described below. The dependent equations, as described above, characterize the interdependency of the binary variables for a targeted deterministic sequence characterized by the essential leading deterministic patterns. Table 1 below shows the listing of this deterministic pattern employed herein as an example, featuring a pattern width N of 6, and a pattern length M of 16:
TABLE 1______________________________________ 6 5 4 3 2 1______________________________________ 1 1 0 0 1 0 1 2 0 0 1 1 1 0 3 1 1 1 1 1 1 4 1 1 1 1 0 0 5 1 1 1 0 1 0 6 1 0 0 0 0 0 7 0 1 1 0 0 0 8 0 0 0 1 1 1 9 0 0 1 0 1 110 1 1 0 0 0 111 1 0 0 1 0 012 1 1 1 1 0 013 0 0 1 0 0 014 1 1 0 0 0 015 0 1 1 1 1 116 1 1 0 1 0 0______________________________________ As described above, an algorithm similar to that shown in Algorithm Listing 1 can be employed to generate a set of dependent equations, and in this example, the set of dependent equations as generated can be listed as follows:
V
V
V
V
V
V The implementation of an algorithm such as that shown in Algorithm Listing 2 can result in a solution that can be employed to construct a sequence generator shown in FIG. 2 of the accompanied drawing. The hardware configuration as outlined in FIG. 2 can then be utilized to generate the pattern sequence as listed in Table 2 that follows. An inspection to Table 2 reveals the fact that the generated pattern sequence contains a leading 16 patterns that conform to the deterministic pattern as listed in Table 1 above. All patterns following the required deterministic patterns are generated pseudo-random patterns. After a cycle of 1,260 patterns in the generated sequence, the deterministic patterns are repeated again.
TABLE 2__________________________________________________________________________1 100101 45 001100 001001 0001002 001110 000000 80 70 001111 0100013 111111 011001 010011 115 0011114 111100 000011 011000 0010005 111010 101000 101010 0111016 100000 50 40 011111 011010 0011107 011000 101111 85 110111 1100108 000111 000100 100010 120 110 0001119 001011 100010 100000 01101010 110001 001101 000011 10111011 100100 55 010010 000010 01000012 111100 000100 90 80 101111 00110013 001000 101001 011001 125 00101014 110000 000101 010100 00001115 011111 011111 000100 01001116 110100 60 50 110110 000110 110000 000001 110110 95 100011 101101 000110 001110 010001 130 120 100100 100101 000100 010001 00011020 100010 010011 011011 110101 111100 65 111000 000001 101111 100111 000000 100 90 010001 010101 010110 111100 001001 135 111101 010001 111111 111011 111001 010011 110110 110000 100010 101100 70 60 100000 110010 000110 001001 110000 105 100101 011010 110000 000000 000111 140 130 001001 111000 110101 000000 01101130 111111 000000 111001 011010 010111 75 110011 101001 010100 011000 110011 110 100 001110 101101 001011 011110 010111 145 010011 100010 001111 111011 000000 001010 100011 000111 100111 101011 110111 100110 000110 010100 001000 100000 011100140 001110 010000 100010 111010 011100 55 190 110101 105 111010 155 110110 110011 010010 240 010100 100110 001001 001001 000011 100011 010010 101101 010000 100101 111011 001001 000011 60 010001 110 101101 160 290 010111 000110 111110 000100 110011 101011 111101 101100 101001 111100 001111 010011 001001150 001001 110110 100111 110110 100010 65 200 100100 115 000001 165 001011 000010 111101 010010 011110 110100 001010 101011 101100 100101 250 100000 001111 000110 101000 011100 001011 111100 70 100110 120 000011 170 300 111000 010000 110111 101110 011011 010000 001110 101100 111001 101110 100011 000010 011011160 010111 011000 111000 011110 75 101110 125 110101 175 010000 101101 210 011110 001101 001000 111010 010010 001000 001110 010101 111111 260 010011 100110 101001 000110 010000 001110 110000 80 000111 130 000111 180 310 010110 101110 111111 000010 101001 000001 010010 001110 011001 010011 001110 100001 111011 111110 110011 101111 100010170 110011 85 010101 135 100110 185 001000 010001 220 000111 111111 000100 000011 000100 100011 111100 100001 111101 270 111011 011110 100000 001100 100001 001110 101010 90 010011 140 101101 190 320 011011 010111 110001 010001 001100 111111 110111 000010 111001 000010 010010 101110 101100 001010 111001180 001011 95 011000 145 100101 195 110101 001001 230 101111 001010 101010 101110 001100 111011 000011 101011 001111 280 011001 000101 100011 010101 001110 001000 100 101011 150 200 101001 000010 010100 011100 001100330 111110 110010 110001 001000 000110 380 010010 111110 110100 100000 000001 430 101011 011011 110100 55 111111 105 110001 155 480 010001 000100 111101 110010 010000 101000 000010 011100 011101 111101 001100 100000 010101 110010 010110 100001 110111 60 101010 110 100111 160 011111 100011 001011 000111 000001340 110010 101000 001110 110000 011110 390 010110 100111 110001 101000 010000 440 011100 000000 000010 65 000001 115 165 010010 011001 110001 001011 490 001011 001101 100011 000011 100101 110110 011101 110100 101101 101010 111000 111011 110000 101010 70 100110 120 111011 170 001111 010000 100101 111010 111001350 101101 110100 101000 101000 110010 400 000010 110101 011100 111000 011111 001010 110001 75 100011 125 450 110010 175 010001 001001 110000 110010 500 101001 101101 101001 101011 010001 001110 100000 011001 010110 110110 100011 110101 111010 000011 80 101000 130 111100 180 110001 111111 001001 111101 000111360 110011 010001 000001 110011 000000 101000 111011 110110 410 000100 010001 110011 001000 85 110101 135 460 110001 185 010111 110111 100001 110001 510 110000 111110 001011 100111 001100 000100 001000 011001 011100 010111 111100 011101 111101 111100 90 110110 140 111011 190 010111 100101 001001 000011 010110 100001 100101 110000 010101370 011001 100011 111111 010111 110010 420 101110 011110 001001 011101 95 111010 145 470 100111 195 010100 001101 111100 001110 520 010110 010111 111100 110001 010010 010011 100010 111100 010001 110101 001111 000010 110010 100 111011 150 100101 200 101010 001111 111111 101111 111100 101100 101001 010100 010111 110011 100001 101111 110001 001101 110001 100100 011110 100111 55 110101 105 001000 155 000001 110101 100011 101011 010111530 011110 111100 011010 111010 100000 580 000000 010000 010110 101111 011101 630 101001 001100 000001 60 000111 110 110110 160 680 100101 110100 001101 011101 110101 000010 101100 100101 100101 001001 101111 011101 001111 000000 101001 010111 101101 65 011111 115 110011 165 000000 100001 110000 001011 101011540 010000 000101 010100 010011 001101 590 100011 011000 101111 011000 111000 640 100010 101100 011110 70 000100 120 170 100101 111001 111000 110111 690 110011 011101 011010 101111 100011 101101 101011 011011 010011 111101 000110 100011 100110 110100 75 110100 125 001100 175 010011 011001 111101 100111 100000550 100110 001000 010010 100000 010100 600 010111 010101 001101 000001 110110 011100 010000 80 010100 130 650 110001 180 011010 011011 100010 101111 700 011100 100110 011111 110100 000000 011010 101000 000101 001011 011111 100101 000010 111001 100101 85 010010 135 000000 185 001001 011110 011011 111111 100100560 110010 101000 011010 110001 000100 100011 110011 010101 610 101010 101011 010000 100000 90 100000 140 660 101100 190 010001 001111 001010 000100 710 111100 101011 111101 001011 001010 010010 101010 011000 000111 111101 001010 110001 010111 110001 95 111010 145 011111 195 010101 001111 001011 010011 110000 011110 111010 001010 000010570 100011 101100 010000 000010 001010 620 110001 111011 001000 010100 100 111010 150 670 001100 200 010010720 000010 111000 100100 110101 000111 011011 110011 011010 770 011101 000011 110001 101000 001111 820 100110 101001 100011 55 100000 105 000001 155 870 000111 100001 011010 011010 101000 110110 110000 000001 111110 010100 000011 110110 010111 010011 010000 000010 111111 101010 60 010101 110 101101 160 000000 111010 111000 101110 110100 011101 101111 011100 010010730 000100 780 011010 100101 101110 101000 110001 830 110111 011001 010111 65 010001 115 100111 165 880 110011 000010 001101 110110 011100 111111 111011 001111 011011 010101 110101 010000 001110 101111 100100 111101 011101 70 000011 120 011101 170 111111 100001 000011 110000 010001 010100 111101 001011 101011740 001001 790 000100 111101 100110 111101 111110 840 001101 011110 000011 75 101001 125 175 011011 011111 000110 000001 890 000000 011001 111010 101100 000011 010000 100011 010101 010100 000000 111101 010100 010001750 100011 80 010010 130 101100 180 011101 111110 000000 000001 110010 110111 101101 011101 111100 010101 800 100000 100000 110101 100110 000110 101100 111100 85 000010 135 850 010100 185 110000 000110 010000 010111 900 101000 101110 110010 101110 110010 101010 110011 011111 111110 110001 111001 000000 000010 010101 90 000100 140 100010 190 100100 001001 011111 101100 010001760 011010 010001 101010 011001 111101 111111 110110 001000 810 011000 001011 011001 001001 95 011010 145 860 111100 195 101000 000111 111101 011111 910 001010 111110 111000 011100 110100 000000 110000 000110 010101 010011 011011 111000 001111 011001 100 110000 150 010100 200 101001 100101 100111 101100 100010 110101 010111 101110 100001 010000 000011 111011 010001 111110 001110 101110 011111 011010 55 000110 105 100001 155 000101920 000100 100100 111001 110100 110000 011000 010100 100000 970 100001 111011 011101 111010 001101 1020 000101 110100 011010 60 111011 110 000011 160 1070 001111 100101 010001 010010 011000 011000 001010 100010 010000 000001 011110 110000 100111 111000 100101 001110 101000 011100 65 111001 115 101011 165 101000 011000 111001 110101 101001930 101011 000011 110110 111011 001000 980 111110 100111 100110 101010 100001 1030 010001 001100 001000 70 101011 120 110000 170 1080 100001 001101 100010 100011 000011 011010 111010 011011 111110 110011 100001 011101 011101 110000 111000 110101 75 100101 125 111011 175 001010 011101 010000 011010 100110940 000101 101100 000010 101111 111010 990 100000 010011 101110 001010 100101 1040 010111 010010 101000 80 011011 130 180 111011 110001 100100 100110 000010 011011 101101 110010 1090 100010 010111 010111 011000 100010 101111 000100 011011 110100 001101 85 101111 135 101111 185 001010 011101 110110 100101 001110950 001000 110010 011010 101000 000001 1000 001010 110110 010000 101010 110000 100011 100110 90 100001 140 1050 111101 190 101100 111001 001110 101001 001101 010000 011110 101111 1100 110100 010110 000100 101111 011101 011101 000001 000000 100010 111010 95 000010 145 100110 195 110001 010001 100001 101000 011100960 101101 000111 001111 110110 100001 100010 011111 010101 1010 000001 101101 011011 110011 100 010010 150 1060 111000 200 100111 001000 110100 100010 1101111110 000010 100110 100110 011011 001111 1160 001101 110000 111000 110011 101111 001001 110010 55 110011 105 1210 010011 155 000011 001100 111100 001110 1260 111010 010010 010010 010100 1 100101 100011 100111 011110 2 001110 100010 100110 010001 010011 60 110110 110 011010 160 000001 011100 0100111120 011000 111110 001100 110010 100011 110110 1170 000011 010110 111110 65 011110 115 1220 11000 100011 001011 101011 011100 001101 010111 101010 101010 101011 111011 001110 001001 001111 70 111110 120 111000 011001 101110 010101 000010 100111 1001111130 111010 001010 101110 000010 1180 110101 011011 000101 75 000111 125 1230 000100 101001 010010 000111 100010 000011 110001 010100 001000 001100 100000 110101 111011 101000 80 001001 130 110110 100000 001100 000001 000100 110110 0010001140 100011 001101 111011 011111 1190 100001 010111 001111 85 000110 135 1240 111001 101110 110011 000100 011100 110011 000101 111000 011001 000110 000001 101110 000100 90 101110 140 111001 011010 100010 011001 000111 011100 1010011150 000101 001101 011000 101111 1200 110000 101001 010001 95 145 111111 111001 011001 1250 100010 011100 000111 101001 111111 101100 111100 100000 111010 000111 011110 100 110010 150 111100__________________________________________________________________________ Thus, for a specifically required deterministic pattern, the present invention makes use of a cascade of inverting non-linear feedback shift registers (namely the inverting non-linear autonomous machine) to constitute a sequence generator capable of generating a sequence of patterns containing a leading deterministic pattern and followed by the pseudo-random patterns of designated length. The constituent registers in the cascade can be selected in accordance with the solution to a set of dependent equations describing the interdependency between the patterns of the deterministic sequence. The utilization of the time-frame expansion scheme (delayed signaling) allows for the set up of the set of dependent equations, and the implementation of schemes such as binary relaxation allows for the resolution of the dependent equations for an optimized set of solution that can be based to construct the cascade. The constructed cascade of shift registers for the sequence generator constitutes an inverting non-linear autonomous machine that is typically characterized by its simplification and regularity in structural configuration. The incorporation of the delay signaling and inverse signaling in the set up of the dependent equations allows easily for the close connection of the number of shift registers in a feedback path of selected XOR gates. This establishment of closely-connected shift registers and XOR gates makes it possible to construct the desired sequence generator based on the minimum possible hardware configuration. As persons skilled in this art can appreciate, the multiple-pattern sequence generator built in accordance with the present invention is suitable for a number of applications. These include the pattern generator for high efficiency built-in self-test sections of VLSI IC devices, coders and decoders as well as compression and decompression sections for digital signal processing equipments. Within the realm of these applications, the sequence generator of the present invention can be constructed to generate test patterns that can verify faults difficult to be pin-pointed by the conventional BIST designs in sequential circuits. With the utilization of the sequence generator of the present invention, memory devices required for the storage of large amount of test patterns can be eliminated to further reduce the overall cost involved in the fault testing procedures. The above-described preferred embodiments of the present invention is utilized only for the purpose of the description of the present invention. Persons skilled in this art can appreciate the fact that other similar arrangements can be devised from the embodiments disclosed above without departing from the spirit of the present invention, which is recited in the claims that follow. Patent Citations
Referenced by
Classifications
Legal Events
Rotate |